General

In principle, activated carbon is nothing more than a collection of pores, the sizes of which are mainly determined by the raw material used and the activation method.

As raw material practically all carbonaceous material can be used; e.g. peanut- and riceshell, bonechar, peat, coconutshells, wood, lignite, coal, anthracite, etc.

The use of coconutshells will provide a very microporous activated carbon which is very much suitable for airpurification purposes, especially in filters with small quantities of activated carbon, such as cigarette filtertips, domestic potable water filterunits, gasmasks, etc.

Wood will give a very macroporous carbon, extremely suitable to remove the larger colour molecules from liquids.

When selecting the right carbon for the job, it is of great importance to understand what the carbon is expected to do: if you choose a type of carbon on the basis of its low price only, you may end up with a much larger carbon consumption than expected and higher replacement costs.

Datasheets

For those people who do not regularly work with activated carbons, it is very difficult to 'read' datasheets. Only some of the testmethods mentioned in datasheets are the same everywhere, described in international standards (DIN, AWWA, ASTM) and used by most carbon manufacturers. However, almost every carbonmanufacturer has developed his own testing methods. This makes it almost impossible to compare grades from different sources. Of course, one can always ask for the description of these methods, but it takes very specialised knowledge to be able to evaluate the figures correctly. Simply comparing values as such may get you offtrack. A very good example is the “molasses-number” which is an important parameter used in the foodindustry. Molasses is a special sugar solution in which almost all types and sizes of colour molecules are present.

Norit The Netherlands defines the 'molasses-number' as the quantity of activated carbon necessary to decolourize a standard quantity of molasses down to a preset level. So, the smaller the amount of carbon required, the lower the molasses-number and the better the activated carbon (for this application!).

Ceca-France uses a fixed amount of carbon to see how much molasses can be treated in order to obtain a preset level. This completely different approach results in higher molasses-numbers for a good quality carbon.
To make matters even more complicated, the introduction of ISO certificates resulted in  making a difference between 'specifications' and 'typicals' on the datasheets.

The specifications are fixed, firm and guaranteed by the manufacturer. These data are tested and analysed on a regular basis. The problem however is that the margins (i.e. the difference between the upper and lower values) are sometimes quite substantial. The typicals are not checked regularly but are average values taken over a few years of production. These values are closer to what is actually supplied.

Forms and shapes of activated carbons

Activated carbons come in two main forms: powder and granular.

The granular carbons can be separated into broken types, extrudates (pellets with fixed diameters) and specials such as spherical carbon. The latter is powdered carbon mixed with a food grade inert material that will desintegrate in water and other liquors.

The definition of powdered carbon is unclear: one manufacturer states that it is all carbon with particles smaller than 80 mesh (0.18 mm). Others set this value at  325 mesh (0.05 mm).

Main applications are the treatment of liquids and the purification of offgases (removal of dioxines and mercury) Powdered carbons are used 'batch-wise' and mostly on a throwaway basis:  simply add the carbon to the medium to be purified, stirr well, filter and throw away the spent carboncake.

Broken type carbons are used in fixed carbonbeds to purify all types of water (potable, process and wastewater) and other liquids. Depending on the hardness of the carbon and the loading, it may be well possible to regenerate the spent carbon effectively. Most common fractions (particle size distributions, see below) are  8x30 mesh (2.36-0.6 mm), 6x16 mesh (3.35-1.0 mm) and 12x40 mesh (1.7-0.425 mm).

Extrudates are made by mixing cokes (a semi-manufactured material) with inert binding material. The resulting paste is pressed through extruders with holes of different diameters, varying from 0,8-10,0 mm. After drying and final activation the spaghetti-strings are broken into smaller pieces. The usual length of the pellets are 3-5 times the diameter. Spent extrudates can also be regenerated.

Iodine number

The iodine number,according to the American Water Work Association (AWWA), is the quantity of iodine (expressed in milligrams) that can be adsorbed on one milligram of activated carbon giving a fixed endvalue.

For the waterpurification industry an often used value to compare the activity of different types of carbon.
For carbon manufacturers it is an important parameter to compare the results of different production runs.

As a practical figure, the iodine number is pretty useless because nowhere people have to remove pure iodine from water!

High iodine numbers do not necessarily mean high efficiency. Coconutcarbons always have very high iodine numbers, due to the raw material. Nevertheless, coconutcarbons are rarely used for waterpurificationpurposes, and certainly not for wastewater.

Hardness

This is a very important value. Activated carbon is always subject to mechanical breakage during packing, transport, filling adsorbers, regeneration, etc. Sorry to say, that here too, carbonmanufacturers use different testing methods to determine the hardness values.

There are three main methods: the abrasion number according to AWWA (the American Waterworks Association), the hardness number according to ASTM (the American Standard Test Methods) and the abrasion index according to own testing methods. The starting point however is always the same: activated carbon is sieved and placed into a rotating perforated cylinder, filled with steel or ceramic balls.

The abrasion number indicates the relation between the diameters of the carbon before and after the test. Mostly used for broken type carbons. The hardness number is determined by the difference in weight of the tested carbon before and after the test. Mostly used for extrudates.

The abrasion index is the difference between the first sievings before and the still remaining sievings after the test. This ratio figure is used to compare the hardness of extrudates only.

Density

There are at least five (!) ways to express the density of activated carbon:

1. the actual density: established by simple test where the carbon is placed into a measuring glass and vibrated for a certain period of time. After that, the weight of the carbon per liter is measured.
   
2. the apparent density: test as above but without vibration. The endvalue is approximately 6% lower.
   
3. the backwashed-and-drained density. Only valid for extrudates and broken types when used in watertreatment applications. Before start-up of the filters, the carbon is subjected to a backwasprocedure during wich the dust and smaller particles are removed and the oxigen in the pores is replaced by water. The carbonbed will expand, so that the weight of carbon per liter becomes less. (To avoid unnecessary carbonlosses, adsorbers are therefore never filled to the rim with dry carbon! A minimum freeboard of at least 20% of the total adsorberheight is advised. Therefore, a filter of 2.00 meters high is filled up to max. 1.60 meters).
   In order to calculate the netweight of the required carbonvolume, the density backwashed and drained is used. This value is approx. 0,88 times the apparent density (see 2). So, an apparent density of 480 gr/ltr results in a backwashed density of 425 gr/ltr. See also Example below.
   
4. the density, wetted in water: with wetted carbon, the oxigen in the pores is replaced by water. This will increase the density to above that of water only. This is the reason why wet carbon will never float!
   
5. the real density (or specific gravity). This is the nettweight of the carbon only.
   Activated carbon consists of about 2/3 open pores, filled with air. The rest is the actual carbon skeleton. The real density is close to that of the raw material itself. For coal this is about 2 kgs per liter.

Ash content

The total ash content is determined by heating the activated carbon to 950 ^oC for a period of 3 hours in an oxidising environment. The carbon itself is burnt away and the remains are ashes. The ash content can be defined as total ash, water soluble ash and acid soluble ash. The ash consists mainly of minerals, already present in the raw material but mostly silicium (sand) and calcium. These are inert, so high ash as such will have no influence on the adsorptive process.

Wood and coconut are by nature low in ash content (< 5%), while coal or peat have 6-10%  ash. In effluents, high ash content can cause a temporary rise in pH value. In the gasfase applications, certain organic components, such as aceton, will be decomposed by high ash contents. For recovery of such expensive solvents this may be a problem.

In the food industry, high ashes are unacceptable. Washing with organic acid, such as phosphoric acid, will remove most of the mineral ash.

Phenol adsorption

Same thing as for iodine. There are three different tests (DIN and 2x AWWA, one  for powder and one for granular carbons) who are all very sensitive to the pH value of the testwater.

Benzene adsorption

Also a parameter commonly used to compare the activity of different types of carbon, mainly for gas phase applications. The practical value however, is higher than for the iodine or phenol values.

Benzene numbers are determined at three different levels of saturation (90, 10 and 1%).This give a clear view on the pore size distribution of the carbon. Standard testgasses such as benzene (C₆H₆) and carbontetrachloride (CTC or CCl₄) are less and less used because of their carcinogetic nature and are more and more replaced by less damaging substances such as butane (C₄H₁₀). The main disadvantage of butane however, is that the testmethod is based on only one saturation point, which in practice does not occur.

To recalculate the CCl₄  value to the butane adsorption number, we use a factor 2.57 (so, a CCl₄ number of 60 mg/g gives a butane ads. cap. of 23.4 mg/g).

The PH value

Most activated carbons have a pH value of >7.
This can be adjusted by (phosphoric) acidwashing or neutralisation with a calciumcarbonate solution, depending on process requirements.

BET surface

This value is determined by a 60-year old method using liquid nitrogen and gives the total surface in square meters of one gram of activated carbon. The test itself is rather expensive and requires very specialised equipment. The result is not an absolute value but a value according to a theoretical formula.

There is a correlation between BET values and Iodine numbers: both figures differ no more than  50-100. If the iodine number is 1.000 mg/g, the BET value will be between 1.050 and 1.100 m2 per gram.

You will also see that high BET values give high benzene adsorption figures.

Pore size distribution

The distribution of the pore sizes (micro, meso, macro) can be established by means of a nitrogen or mercury adsorption test. The pore size distribution is one of the parameters determining the efficiency of a certain type of carbon for the application.

The adsorption technique is as follows: the macropores (> 50 nm) create easy access of the molecules to the smaller pores. The mesopores (2-50 nm) influence the speed of adsorption and the transport of the molecules into the micropores (<2 nm) . Here the actual adsorption takes place.

If there are too many macropores in relation to the smaller pores, components of small molecular size will not be adsorbed efficiently and the carbon is not used effectively.

Halving value

The removal of chlorine or ozone from water is no adsorption technique but a combination of a catalytic and a chemical reaction that takes place on the surface of the carbon.

In principle this is an endless process that can only be influenced (shortened) by the presence of dissolved organic matter. Also an excessive dosage of free chlorine or ozone can 'burn away' the very thin carbonwalls between the pores.

The halvingvalue is defined  as the number of centimeters necessary to halve the initial amount of chlorine in water at a velocity of 36 m/hr and 20 degrees C.

Pressure drop

The pressure drop over a carbonbed is the difference between the incoming and outgoing pressure caused by the resistance in the carbonbed. The height of the pressuredrop is influenced by a.o. the carbon bedheight, the diameter or particle size of the carbon, the air- or watervelocity, the viscosity of the medium, etc.

The theoretical correlation between velocities and carbonbedheights is usually shown in a pressuredrop graph on the datasheets. This is the only parameter you will find, that has no relation whatsoever to the quality of the carbon but is the result of the physical form and shape of the carbon.

Moisture content

The total amount of moisture adsorbed by the carbon is determined by heating a sample to 150 ^oC for a period of 3 hours and measuring the difference in carbonweight before and after the test. The value found is expressed as a weightpercentage. Normally, broken types and extrudates have a moisture content of 2-5 wt%. Powdered and impregnated carbons show higher values, as do acid- or waterwashed grades.

Filtration velocity

This value is mainly of interest when using activated carbon in liquid phase applications.A higher amount of very small particles will fill the space between the the larger particles and increase the risk of blocking the carbonbed. The result will be a decrease in filtration velocity and a higher pressure drop over the carbonbed.

Datasheets will show a number of figures to explain the filtration characteristiques of the carbon:

1. the effective particle size: a figure of 0.7 indicates that 10% of the total is smaller than 0.7 mm.
   
2. the uniformity-coëfficiënt is the ratio between the diameter at which 60% of the total is smaller, and the effective particle size. If 60% is smaller than 1.0 mm and 10% is smaller than 0.7 mm, then the UC = 1.0 ÷ 0.7 = 1.4.
   At a Uniformity Coefficient of 1.0 all carbonparticles have the same size (this is only the case with extrudates which have fixed diameters). The smaller the UC, the uniformer the activated carbon and the higher the filtration velocity.
   
3. the average diameter. This value is mainly of interest in gasphase applications. It influences the pressure drop over the carbon (see above).

Regeneration

Powdered carbons are rarely regenerated but usually dumped, burnt or used as fuel for kilns.    In those cases where the powdered carbon has a purely organic loading, it can also be used as soil improvement or even mixed with cattlefeed.

In priciple, only broken type carbons and extrudates of sufficient hardness are suitable for regeneration. Usually, regeneration is handled by the carbonmanufacturers themselves. They are bound by government regulations that mention criteria for acceptance.

The regeneration or reactivation process is similar to the activation process. The spent carbon is heated to 1.000 degrees C and the components with lower boilingpoints are removed from the carbon. During this process however, the carbon skeleton is affected: and the walls between the micropores are gradually burned away so that the BET value decreases, the poresize distribution is changed and there will be carbonlosses (about 10-15%).
Depending on the type of loading, the adsorption capacity  after regeneration will be 80-100% of the original value.

Physical qualities of activated carbon (such as hardness, densities, etc.) are values that can be of importance for the process. Low density carbons are often promoted by their manufacturers because of the seemingly price advantage. Adsorbers with a fixed volume will indeed contain less kilogrammes of lightweight carbons. However, as the adsorption capacity is always expressed as a percentage by weight,  such carbons will have a lower total adsorption result and will have to be replaced more often.

Adsorptive properties such as iodine , phenol or benzene numbers should be jugded with extreme care. These values are determined under very strict and controlled laboratory circumstances, that bear no relation to practical conditions. There we usually see a mix of components with different values, that vary in time. The saturation point of the carbon is then influenced by the one component having the least affinity with activated carbon. Whether or not there may still be adsorption capacity left for the other components is of no interest anymore. A reliable basis for determining the lifetime of a carbonfilter is simple: only pilottests that approach the actual conditions of the process will give you usefull information.